Deciding whether to robotically inspect a generator is a big decision. One reason is cost: It’s expensive, typically running more than six figures. Another is accessibility: Only certain types of generators have air gaps large enough to allow robot access. Yet another: You might not get all the information you want.
However, robotic inspections do provide some useful information and they take only three or four days to complete (photos). This is much less time than required to pull the field, conduct an inspection, and replace the rotor—and without exposing the generator to the possibility of damage during handling.
How would you decide between robotic or manual inspection? Consulting Editor Patricia Irwin, PE, asked two plant owners, in very different situations, this question, and their answers were inconclusive: Robotic inspections can be useful, in certain cases, the editors were told. Careful evaluation of your particular situation will determine the best way to proceed.
Duke Energy
Duke owns and operates more than 350 generators. With that number of machines, Kent Smith, manager of generation engineering, put a great deal of thought into whether Duke should buy its own robot. There are several suitable robots on the market, he said, ranging in height from half an inch to two inches. They typically are 6 in. wide and about a foot long. Each has a drive mechanism, which takes up the majority of the available space on the device, and some magnetic properties, so the robot can hold itself to the top of the core. Different sensors and modules are attached to perform specific tests.
Smith evaluated several robots and found that size does matter. “Robots can only be used in certain types of machines because of the narrow air gap. That is true even for OEM robots on their own generators. For example, the Siemens robot can’t inspect all Siemens machines,” explains Smith. “So, we evaluated about 160 of our machines and determined how many we could inspect robotically. The answer was only about 10% and we decided it really wasn’t worth the effort or money to buy our own robot.”
But Duke does hire OEMs to inspect some of the generators that fall into the 10%. For example, Siemens robotically inspects its Aeropacs and, on one occasion, Duke hired GE to use its Miniature Air Gap Inspection Crawler (MAGIC) system on one of its machines. Smith explains, “We don’t tend to robotically inspect our large GE machines because they have air baffles. Basically, the air baffle is a rubber insert, perpendicular to the length of the core, which goes 80% of the way around the core. There can be eight or 10 baffles down the length of the stator core. So, you have this small space down at the bottom of the core, near the skid plate, to get the robot out. To see anything else you have to run the robot down to the skid plate and then try to run it back up the side. And, the robots just don’t move that way.”
Types of tests. Assuming the robot fits inside the air gap, there are other things to consider before scheduling an inspection. “Decide what you want tested and if the robot will give you the kind of results you need. For example, you can test wedge tightness with a ‘tapping’ module, you can do an Electromagnetic Core Imperfection Detection (El CID) test, and you perform a visual inspection. But, at Duke, we are going back to running loop tests on our cores and you can’t do that with a robot. So, if we are going to pull the field anyway, why spend the $100,000 for a robotic inspection?”
Missed spots. Since the robot cannot reach everywhere, spots will be missed during the inspection. Smith continues, “The robot doesn’t do a good job on the step iron. To get into the core, it has to get past the step iron and it can’t read the step iron very well. So, if you want to run an El CID at the step iron, you have to do it manually.”
Sticking point. And, while rare, Smith is concerned the robot could get stuck. “We haven’t had a problem but we have heard of it happening. And, we heard the vendor had to do some ‘unique manipulations,’ but was able to get the robot out without having to pull the field.”
Wrap-up. Despite the limitations, Duke decided to continue robotic inspections on appropriate generators for general inspection. “With robotic inspections, there is less disassembly and less risk. Does it give you a complete picture of the health of the machine? No. But, it is a lot better than a normal crawl-through inspection,” says Smith. He continues, “I am a proponent of having more tools in the toolbox. Nothing can be used 100% of the time and give you a 100% confidence level on the health of the machine, so you need to have as many tools as possible at your disposal. Robotic inspections are one more tool.”
Griffith Energy
Compared to Duke, Star West Generation LLC is a much smaller operation, owning only five powerplants. Mike Hartsig, plant manager for the company’s Griffith Energy, has a 2 × 1 F-class combined cycle equipped with two GE 7FH2 generators for the gas turbines and a Toshiba generator for the steam turbine.
Hartsig notes, “In 2012, we had a contractor to do some partial-discharge testing our generators. Conclusion was that we should take a look at the two 7FH2 generators, especially the wedges. We had a four-week outage coming up, so we had time to pull the rotors and physically check the wedges, if we choose to. Our other option was to check the wedges robotically, which we had done in the past.”
In 2005, the plant hired GE to perform robotic inspections on those units and Hartsig decided to use robots again to avoid the risks associated with pulling the rotor. This time around (2013), the plant manager got quotes from various vendors and settled on an experienced vendor, other than GE, to perform the inspection. After testing, the vendor prepared wedge maps for both machines. The color coded maps showed significant wedge looseness, particularly on Unit 2. Vendor engineers recommended a total re-wedge, but Hartsig was not sure.
Contradictory results. “Both machines had been inspected by GE in 2005. So, I had GE come back in with the Magic system and do a sample wedge tightness test on Unit 2. The results? No problem at all. Then, I compared the latest GE results with previous results. In 2005, GE said that the generator looked good but one wedge was questionable. According to the new test, that wedge was no longer a problem,” says Hartsig.
At this point, Hartsig and his team reconsidered the value of robotic testing. “We looked at our original premise and, after much discussion, concluded that our original premise was sound. Even though the results raised a lot of questions, we avoided the risks of pulling the rotor and we still felt comfortable in our decision to do the testing.”
Decide what to do. Griffith Energy brought in independent experts to review all of the wedge-tightness test results. They concluded there was no significant problem. “This was based on their experience and knowledge of wedging in generators. The experts pointed out that, even if a wedge came loose, which was unlikely based on collected data, the worst-case scenario was some arcing between stator bars. And according to the experts, you can run a generator for a prolonged period of time with a shorted bar, although this is not recommended. The general conclusion was to monitor key indicators and wait until 2018 to go into the units,” explains Hartsig.
Conclusion. Hartsig continues to recommend robotic inspections, with a few caveats. “Why do a robotic inspection if you get a contradictory results and you don’t do anything with them? I think those results pushed me to the next level. My advice is to go in with your eyes open. Get the results and compare them to other data you’ve collected. Then make the hard decisions for yourself.”
Hartsig also recommends relying on industry experts who will not gain anything by recommending you wait before doing remediation work. “A larger company will have a massive back office with its own experts and engineers. We don’t, so we go to outside contractors. And, I don’t take the word of one contractor, I get a second or third opinion. Then, based on that, we move forward.”
“I still favor robotic inspections for one simple reason,” he said: “The risk of damage is slim to none. The risk of damage from pulling the rotor is greater. That is not to say that every time you pull the rotor you’re going to do damage, but the risk is always there,” concludes Hartsig.
